Databricks workspaces in an account

Lists every workspace in a Databricks account using the provider's built-in vw_workspaces view. Prefer the vw_ views over querying the underlying resources directly: they are shipped inside the provider with per-dialect implementations, so the same query works on the embedded SQLite backend and on a PostgreSQL-backed instance without rewriting any JSON extraction.

Query

SELECT
workspace_id,
workspace_name,
workspace_status,
cloud,
aws_region,
location,
pricing_tier
FROM databricks_account.provisioning.vw_workspaces
WHERE account_id = '{{account_id}}';

Creating a workspace

A workspace needs a credential and a storage configuration to exist first - see databricks/provisioning/account-infrastructure for both:

INSERT INTO databricks_account.provisioning.workspaces (
  aws_region,
  credentials_id,
  pricing_tier,
  storage_configuration_id,
  workspace_name,
  account_id
)
SELECT
  'us-east-1',
  '{{credentials_id}}',
  'PREMIUM',
  '{{storage_configuration_id}}',
  'my-workspace',
  '{{account_id}}';

Notes

The region column that carries a value depends on the cloud: aws_region is populated for AWS-hosted workspaces while location carries it for Azure and GCP, and cloud itself is often null on AWS accounts, so treat the presence of aws_region as the AWS signal rather than testing cloud. workspace_id is a numeric id, not a GUID, and it is the key for every per-workspace query - including the account-level assignment views. Creation is asynchronous: workspace_status moves through PROVISIONING to RUNNING, so poll this view rather than assuming the insert completed the workspace. Per-workspace resources such as clusters, jobs and catalogs live in the separate databricks_workspace provider, which authenticates against the workspace host rather than the account, so an account-level credential alone will not reach them.